The long-term objective is to construct Salmonella-based anti-HIV vaccine strains for trial in volunteers, consisting of an aromatic-dependent strain of S. typhi (needing p-aminobenzoate, not available in host tissues, therefore non-pathogenic) carrying genetic information such that its administration will cause humoral and cellular immune responses to both the bacterial carrier and to the HIV epitope encoded by the passenger DNA. The foreign DNA will be a short synthetic oligonucleotide specifying an amino acid sequence identified as an epitope of an HIV protein antigen able to cause production of virus-neutralizing antibody or other protective response. The oligo will be inserted in a cloned flagellin gene at a point such that the amino acid sequence it specifies is exposed at the flagellar surface, accessible to antibody and expected to be highly immunogenic. Another form of the vaccine will consist of flagella broken off from the live-vaccine bacteria, concentrated and purified for use as a product vaccine, to be given by injection. Specific Aim 1 is to test immune responses of mice and guinea pigs to parenteral or oral administration of a Salmonella live vaccine with flagella made of flagellin specified by a cloned gene with oligos determining short amino acid sequences from gp160 of HIV inserted at the EcoRV site, by use of already constructed or to be made derivatives of plasmid pLS408 placed in a flagellin-negative aroA live-vaccine strain of S. dublin. Aim 2 is development of a method for quantitation of flagellin and the investigation of what features of an amino acid insert determine continued production of flagella or the accumulation or secretion of flagellin without production of flagella; and of genetic methods of increasing amount of flagellin produced by a strain and the effect on immunogenicity, etc, of having two or more inserts, specifying the same or different epitopes, in tandem in a single flagellin molecule. Aim 3 is the construction of a replacement for a live-vaccine strain of S. typhi already tested in volunteers by one attenuated by two deletions in the aro (aromatic biosynthesis) pathway, instead of by an aroA deletion and a purA deletion, and investigation of effect of variation of parameters of dosage, presentation, etc, on efficacy of orally administered live vaccine.